PROJECT SUMMARY/ABSTRACT Where a child grows up can in?uence his or her risk of problematic drug use. Surprisingly, policies and programs designed to improve neighborhood conditions have yielded con?icting results in which some children bene?t but others are harmed. Understanding the mechanisms by which neighborhoods act to affect problem- atic drug use and how these mechanisms differ across subgroups and cities?that is, understanding where and for whom mechanisms apply?is essential to optimize the effectiveness of future neighborhood interventions. Mechanisms involving mediation by aspects of the school environment have been unexamined, but could prove promising for understanding differential effects of neighborhood on problematic drug use. The objective of this K99/R00 is to identify school-based mechanisms that mediate the relationship between neighborhood poverty and problematic drug use and to develop statistical methods to understand how mediation effects differ across subgroup and place. This objective contributes to the long-term goal of improving the effectiveness of interven- tions to prevent drug abuse and dependence by tailoring interventions to address the most relevant mechanisms of action for each target population based on place and individual-level characteristics. The K99 phase ?lls the following training gaps that are critical to achieving the research objective and long-term goal: 1) methodologic training in causal mediation analysis; 2) subject-matter training in drug abuse and dependence prevention in ur- ban populations; and 3) involvement in real-world school-based interventions to develop methods with practical utility. Aim 1 of the proposed research identi?es the school environment mechanisms that mediate the relationship between neighborhood poverty and adolescent problematic drug use and modi?ers of those mechanisms. Aim 2 extends a statistical method recently developed by the proposed research team for generalizing intervention ef- fects from one city to another to generalizing mediation effects. Finally, Aim 3 uses the method developed in Aim 2 to identify how the school environment mechanisms mediating neighborhood poverty and problematic drug use differ across major U.S. cities. The proposed research is expected to identify school-based mechanisms underly- ing differential effects of neighborhood poverty on adolescent problematic drug use by subgroup and place using the only study of randomized neighborhood residence. In addition, it is expected to contribute a novel statistical method that incorporates cutting-edge machine learning techniques to identify which mediation mechanisms can generalize across place, advancing translational research. Successful interventions are assumed to work in differ- ent settings, but that is not the case in practice. The proposed research will allow us to predict intervention effects while accounting for differences in population composition and modi?ers of the mediation mechanism. This will make a signi?cant contribution to improving health by informing the design and targeting of future neighborhood interventions to adolescents and places that may be most impacted in terms of problematic drug use prevention.